INDEX

Explanations

complex sequences of special characters and letters likely with a specific language-based pattern

oai_token-act-pair · gpt-3.5-turbo

specific numeric values or sequences possibly related to measurements or ratings

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nesota

-1.15

apy

-0.91

isters

-0.89

icro

-0.86

ept

-0.81

ister

-0.79

oan

-0.78

llular

-0.77

utical

-0.76

ade

-0.74

POSITIVE LOGITS

ãĥ¼ãĥ

0.70

tin

0.67

POSE

0.67

rier

0.66

headed

0.65

Ñĭ

0.65

posed

0.63

vig

0.63

ground

0.63

cv

0.63

Activations Density 0.040%

complex sequences of special characters and letters likely with a specific language-based pattern

specific numeric values or sequences possibly related to measurements or ratings

No Comments

No Known Activations